Intensifying Adverbial Particle or Emphatic Prefix?

24 Pages·2014·0.31 MB·English

Intensifying Adverbial Particle or Emphatic
Prefix? A Study of Intensive Adjectives in
Mongolic Languages*
Hyung-Soo Kim**
1)
- CONTENTS -
1. Introduction
2. Intensifying adverbial particle vs. emphatic prefix
3. Linking consonant(s) in Mongolic languages
3.1 Dag. čim čigaan ‘very white’
3.2 Khlk. bas bat ‘very firm’
3.3 Klm. bim bitü ‘firmly closed’
3.4 Other exotic linkers
4. Conclusion
언어학 제 70 호 (2014. 12. 30: 305-328), 사단법인 한국언어학회
*This paper was written while the author was a visiting scholar at the department of
linguistics, Seoul National University (Mar.-Aug. 2014). I would like to thank the chair
of the department, Prof. Jongho Jun, and my host, Prof. Juwon Kim, for making the visit
possible on short notice. I am also grateful to the three reviewers of the journal and the
following colleagues whose comments and materials sent made the final version much
improved (in alphabetical order): Juwon Kim, Dongho Ko, Yong-Sŏng Li, Andrew
Shimunek, Jae-mog Song, Kamil Stachowski, Maria Tolskaya, Wonsoo Yu. Needless to
mention, I alone take responsibility for any errors.
**Dept. of English Education, Jeonju University
306 Hyung-Soo Kim
1. Introduction
A perennial problem in historical linguistics is lack of reliable data, a problem an
analyst often encounters even in a relatively well-documented language family such
as Indo-European. So the problem is all the more pronounced in Altaic where the
data available is much more limited. It is with this apprehension that I begin my
investigation of intensive (emphatic) adjectives in Mongolic languages. My aim in
this paper is to trace the development of emphatic adjectives in Mongolic, especially
the origins of linking consonants, through comparison with other Altaic subgroups
such as, for example, Turkic languages, but I found my analysis often thwarted by
lack of etymological information, rendering it highly conjectural. In some cases I had
to content myself only with adumbration of possible solutions, with the choice of the
final analysis deferred until a further insight in etymology comes to light. While this
emphasis on etymology is typical in most historical comparative analyses, it seems
all the more justified in Altaic linguistics as there are still plenty of rooms for
exploration, especially of subjects such as intensive adjectives, as is illustrated in this
paper.
References on Mongolic languages perused for intensive adjectives raise a number
of interesting questions on their development, most of which unfortunately must
remain unanswered at this point, due to lack of understanding of the phenomenon
itself and/or paucity of reliable etymological data.1) Of these, I have chosen the
following two for discussion in this paper as they are not only the kind of problems
one typically encounters in analyzing intensive adjectives in Mongolic but are the
ones to which I could suggest some reasonable solutions:
1) Poppe (1954: 59) says that the prefixal part of the intensive
adjectives actually functions as a modal adverb meaning ‘completely’.
But this would mean that the emphatic prefixes are independent
words formed by reduplication of the first (C)V of the corresponding
1) For the type of questions raised, see Kim (2013c), from which this paper has been
expanded.
Intensifying Adverbial Particle or Emphatic Prefix? 307
A Study of Intensive Adjectives in Mongolic Languages
base adjective and appendage of a linking consonant /b/. Such word
formation, however, would be very unusual. There are, on the other
hand, some forms that seem to have the same function as emphatic
prefixes, referred to as ‘intensifying adverbial particles’ by Janhunen
(2012: 213). Traditionally the two terms have been used
interchangeably in Altaic linguistics but making the distinction
between them is a crucial first step for properly understanding the
formation of intensive adjectives in Altaic languages.
2) Janhunen (2003: 12) assumes that the Common Mongolic (CM)
reduplicative pattern is (C)V.b&(C)V-, e.g. *xulaxan ‘red’:
*xu.b&xulaxan ‘reddish, quite red’.2) But some have claimed that
there are linking consonants other than /b/ in Mongolic. For example,
Hugjiltu (1998: 213), Tsumagari (2003: 135) and Wu (1996: 20) all
claim /m/ to be a linking consonant in Dag. čim čigaan ‘very
white’, while Ramstedt (1952/1985: 263) says that Khlk. bas bat
‘very firm’ and Klm. bim bitü ‘firmly closed’ use linking /s/ and /m/
respectively. So the question naturally arises: what are the possible
linking consonants in Mongolic?
In what follows then, I will elaborate on these questions in two sections: intensifying
adverbial particle vs. emphatic prefix (2) and linking consonants in Mongolic (3).
The conclusion highlights the ramifications of the analysis.
2. Intensifying adverbial particle vs. emphatic prefix
The need for this distinction is well demonstrated by the following Middle
Mongolian data from the Dictionary of Sonom Gara's Erdeni-yin Sang:
2) Following abbreviations are used throughout the paper for names of languages: PM:
Proto-Mongolic; CM: Common Mongolic; WM: Written Mongolian; MM: Middle Mongol;
Mo.: Mongolic; Bon.: Bonan; Dag.: Dagur; Khlk.: Khalkha; Klm.: Kalmuck; Khor.:
Khorchin; Mong.: Monguor(Tu); San.: Santa; OT: Old Turkic; AZ: Azeri; Chag.:
Chagatai; Chv.: Chuvash; Tksh.: Turkish; Trkm.: Trukmen; Kir.: Kirghiz; Kaz.: Khazak;
Yak.: Yakut; Xib.: Xibe; Orq.: Oroqen; Sol.: Solon. ME: Middle English.
308 Hyung-Soo Kim
(1) Instensifier/intensifying prefixes in Middle Mongol (Kara &
Kiripolská 2009)
AB² intensifier prefix ab ali čimeg… whatsoever ornaments … (p.2)
QAB² intensifying prefix, qab qamiɣ-a ber … wherever, anywhere… (p.222)
ENG intensifier prefix in… eng urida … at first, first of all… (p.99)
Even though all three entries function as intensifier of the following adjective,
there is an important morphological difference between them: the first two are
formed by reduplication, but the last is not. Yet the dictionary labels them all
‘intensifier/intensifying prefix’. It is clear, however, that ab and qab are emphatic
prefixes but eng an adverbial particle. This is because the latter meaning ‘very,
most’ is nonreduplicative and is an independent word that modifies a following
adjective/adverb as in, e.g. WM eng sacaɣu ‘identical, the same’, WM eng terigyn
‘first of all, very first’ (cf. Lessing 1960), while the former is reduplicative with
linking /b/ appended.
Grouping what are intensifying adverbial particles with emphatic prefixes under the
same category has been a long tradition in Altaic linguistics. Poppe (1954: 57), for
example, writes that:
“There are modal adverbs with the meaning “completely,” derived by
reduplication of the first syllable of the word with the inserted consonant
–b. If the first syllable of the word concerned is no, the adverb is nob;
if the first syllable is qa, the adverb is qab, and so on. qab qara
completely black, qab qarangɣui̯ pitch black, … nob noɣuɣan grass
green…”
He is thus assuming that qab and nob in the examples he provides derive from the
base adjective by partial reduplication but function as independent adverbs that
modify following adjectives/adverbs, just like the Middle Mongol intensifying adverb
eng above.
On the Turkic side, a number of names refer to emphatic prefixes in Clauson
(1972): ‘reduplicative intensifying prefix, usually placed before N./A.’s connoting
colour or physical shape’ (p.3), e.g. OT ap ak ‘intensely white’ and OT ep edgü
‘very good’; ‘an intensifying particle’ (p.709), e.g. OT köm kök ‘dark dust-coloured’;
Intensifying Adverbial Particle or Emphatic Prefix? 309
A Study of Intensive Adjectives in Mongolic Languages
‘an alliterative prefix to Adjs, mainly of colour, indicating intensity’ (p.578), e.g. OT
kap kara ‘quite dark’, OT kıp kızıl ‘bright red’ and OT kıp kırmızı ‘pure red’. But
these are all emphatic prefixes formed by reduplication of the first syllable of the
base with addition of a linking /p/ or /m/. On the other hand, the same dictionary
calls the adverb eŋ, e.g. OT eŋ ilki ‘first of all’ (cf. MM eng above) ‘an Adjectival
Prefix forming a quasi-Superlative’ (p.166) while citing süm in OT süm süçıg ‘very
sweet’ as ‘alliterative prefixes’ (p.828), and tüg meaning ‘several, many’ in OT tüg
tümen ‘several thousand’ an ‘alliterative jingle’ (p.476).3)
In this paper I only distinguish between an emphatic prefix, which occurs when
the adjective is formed by (C)V-reduplication of the base with addition of a linking
consonant, and an intensifying adverbial particle, which typically exhibits no
reduplicative morphology. The former, having been attached as a prefix by affixation,
is not an independent word in origin, while the latter, having originated as an
independent word that modifies following adjective/adverb, is an adverb with
intensive meaning but came to be fixed to occur only in certain adjectival phrases,
which is why it is called an ‘adverbial particle’ rather than a simple ‘adverb’.
To see that the problem is not just terminological but deserves a careful scrutiny,
consider the following intensive adjectives in Khalkha:
(2) Intensive adjectives in Khalkha:
cev cever ‘perfectly clean’
cev celmeg ‘clear and bright’
cev ceŋker ‘quite blue’
cev xüiten ‘quite cold’
Both Hangin (1986: 768) and Bawden (1997:514) regard cev in these examples as
‘intensifying (adverbial) particle’. Lessing (1960: 167), who cites only WM ceb cegen
‘entirely white or light’ and WM ceb cengker ‘very light blue’, on the other hand,
describes ceb as an ‘intensifying particle used before certain adjectives and adverbs
beginning with the syllable ce’, a description some dictionaries still use to refer to
3) Clauson’s (2002: 227) remark on Khlk. tüg tümen as ‘a Turkish-looking reduplicated
form’ suggests that he viewed tüg here to be an emphatic prefix, with /g/ as a linking
consonant. Vide infra (3. 4) for explanation of this and other forms that ‘appear’ to be
reduplicative with exotic linking consonants.
310 Hyung-Soo Kim
an emphatic prefix.
Since they all show reduplicative morphology with the usual default linker /v/ in
Khalkha (from CM /b/), the first three examples are emphatic prefixes under
Lessing’s description (though he calls them ‘intensifying particle’). However, the last
example, Khlk. cev xüiten, is not reduplicative though it occurs with the same prefix,
contradicting our initial supposition.4) The question is then what should we call cev
in these examples: emphatic prefix or intensifying adverbial particle?
The answer depends on the origin of cev. If it turns out to have been an
independent adverb historically, it is an intensifying adverbial particle, and even
though cev in the first three examples happens to resemble an emphatic prefix, it is
actually an adverbial particle, just as in the last example. According to the
etymological information currently available, however, there is no evidence for cev
ever having been an independent adverb with emphatic meaning, neither in Mongolic
nor in Turkic. It is thus likely that cev in these examples originated as emphatic
prefixes by CV-reduplication and appendage of /v/. The fact that examples with ceb
in Written Mongolian as described by Lessing (1960) show the same reduplication
also supports this view. The problem, however, is how we should explain the cev in
Khlk. cev xüiten, which we cannot possibly imagine having been formed by
reduplication.
One way to explain the oddity in Khlk. cev xüiten is by making use of the
knowledge made available from previous study of language change, from which we
know that new forms sometimes arise by ‘reanalysis’. A well-known example is
modern English pea, which was formed when speakers reanalyzed ME sg. pease as
a plural form: /pi-z/. Perhaps the same is happening with cev: speakers, being aware
that there are many intensifying adverbial particles such as WM cel ‘absolutely, very,
completely’ that come before some adjectives and modify them, e.g. (Lessing 1960:
171; Hangin 1986: 769)
(3) Examples of adverbial particle WM cel ‘absolutely, very, completely’
cel køke ‘deep blue’
cel kyiten ‘very cold’
cel noɣuɣan ‘deep green’
4) The base has its own intensive form with the emphatic prefix xüb: Khlk. xüb xüiten;
Khor. [kyp xyytʰən]. Cf. Janhunen (2012: 121) and Svantesson et al. (2005: 59).
Intensifying Adverbial Particle or Emphatic Prefix? 311
A Study of Intensive Adjectives in Mongolic Languages
cel zalaɣu ‘very young’
reanalyzed the emphatic prefix cev in the first three examples as an intensifying
adverbial particle and ‘applied’ it to the adjective Khlk. xuiten, even though it cannot
occur with this adjective if it was indeed an emphatic prefix formed by
CV-reduplication of a base adjective with linking /v/.
The reanalysis just described is often the first step in the grammaticalization
routine (cf. Hopper & Traugott 1993: 49) and here we can say that cev, originally a
reduplicated prefix derived from the Common Mongolic intensive formation of
(C)V.b&(C)V, has been reinterpreted as an intensifying adverbial particle that
modifies an adjective that does not necessarily share the same base. It is notable
that if this process is carried through, that is, if all emphatic prefixes become
intensifying adverbial particles by reanalysis, the distinction between the two will be
lost and what prevails synchronically will be what Stachowski (2014: 204) calls
‘emancipated reduplicated anlaut,’ in which the reduplicated anlaut plus the closing
consonant, i.e. our emphatic prefix, is promoted to an independent intensifier.
Presumably the emphatic prefix cev in Khalkha is in the incipient stage of this
grammaticalization process.5)
It is understandable then why previous scholarship has so often made no
distinction between an emphatic prefix formed by CV-reduplication with addition of
a linking consonant and an intensifying adverbial particle that often occurs before
certain adjectives and modifies them as an independent adverbial. Maintaining the
distinction, however, is important as we will see immediately below when we
analyze possible linking consonants in Mongolic. Previously čim in Dag. čim čigaan
has been regarded as an example of emphatic prefix with linking /m/, but this will
complicate the Mongolic reduplicative pattern of intensive formation in which the
5) We should also keep in mind that there is still another possibility, albeit a remote one:
only the cev in the last example is an intensifying adverbial particle, whose form
‘happens to’ be homophonous with the emphatic prefix cev in the first three examples.
As in the preceding explanation, the viability of this hypothesis ultimately rests on the
etymology of cev. We should also note the abnormality of the first example, Khlk. cev
cever ‘perfectly clean,’ which looks as if it is repeating the first CVC rather than the
CV of the base. Such partial reduplication in emphatic adjectives, though rare in Altaic,
occurs in Chuvash, e.g. Chv. tak-takar ‘absolutely flat’ and is most likely to be due to
‘full-to-partial’ reduction of reduplicative compounds (Kim 2009).
312 Hyung-Soo Kim
linking consonant is typically /b/. An alternative analysis, to be presented below,
considers it to be an intensifying adverbial particle that happens to resemble an
emphatic prefix.
3. Linking consonant(s) in Mongolic languages
It is generally agreed that unlike in Turkic where the linking consonant varies
(among as many as four consonants, e.g. /p, m, s, r/ in Turkish), only one linking
consonant /b/ occurs in Mongolic languages. This is why Janhunen (2003: 12) posits
the CM reduplicative pattern for intensive adjectives as (C)V.b&(C)V-. But there
appear to be exceptions to this general rule, the most conspicuous of which are:
Dag. čim čigaan ‘very white’, Khlk. bas bat ‘very firm’ and Klm. bim bitü ‘firmly
closed’ (Ramstedt 1952/1985: 263). These examples seem to suggest that as in some
Turkic languages /m/ or /s/ could also be a linking consonant in Mongolic, even if
only rarely. In what follows I examine these claims to see if they constitute genuine
cases of linking /m/ and /s/.
3.1 Dag. čim čigaan ‘very white’
Hugjiltu (1998: 213), Tsumagari (2003: 135) and Wu (1996: 20) mention Dag. čim
čigaan as exhibiting linking /m/. There are however a number of problems with this
claim. First, as in other Mongolic languages /b/ is the usual linking consonant in
Dagur: Dag. xub xula:n ‘very red’; Dag. xab xar ‘very black’; Dag. dab dasu:n
‘very sweet’, etc., and Dag. čim čigaan ‘very white,’ if it indeed is an intensive
adjective, is the only exception, with /m/ appearing in place of /b/. Moreover, the
examples that appear in Martin (1961) show no linking /m/. Consider:
(4) Intensive adjectives in Dagur
Adjective Intensive adjective
čigaan čabe-čagan ‘snow white’
hare habe-hare ‘jet black’
šari šabe-šari ‘vivid yellow’
Intensifying Adverbial Particle or Emphatic Prefix? 313
A Study of Intensive Adjectives in Mongolic Languages
adili ab(e)-adili ‘the very same’
amere.han ab(e)-amere.han ‘most easy’
These examples are quite different from the data given by Hugjiltu (1998),
Tsumagari (2003) and Wu (1996). Notable is the vowel /e/ that appears after the
linking /b/, which optionally disappears before another vowel. Martin’s (1961: 110)
entry in his lexicon of ‘abe/Cabe’ as the ‘intensive prefix,’ under which the above
examples appear, makes it clear that at least in the dialect he is describing /b/
succeeded by an optional /e/ is the regular formative marking intensive adjectives.
Thus what is peculiar is not the intensive adjective Dag. čabe-čagan but the lack of
correspondence between the base Dag. čigaan, and it’s intensive adjective, which
should have been the incorrect *čibe-čigaan. Since Dagur has many dialects due to a
complicated ethnic history, this may be due to a dialect mixture, or contamination,
perhaps from the standard dialect.6)
The foregoing discussion indicates that to claim Dag. čim čigaan to be a genuine
case of intensive adjective, one should explain why this particular adjective uses /m/
for its linker rather than the default /b/, which seems to be used with all other base
adjectives in Dagur as well as in other Mongolic languages.7) To understand the
logic behind this insistence on explanation, we may draw on a similar case from
some Tungusic languages. The linking consonant in Xibe is, as in Mongolic,
uniformly /b/, which sometimes appears as /v/ before voiced consonants, as in, e.g.
Xib. gov golmin ‘very long’; Xib. nav narxun ‘very thin’; Xib. xab xalxun ‘very
hot’, etc. But as Hugjiltu (1998: 221) notes, occasionally words that begin with
labial consonants add /q/ instead, as in Xib. faq farxun ‘very dark’. The same holds
in Oroqen (Hugjiltu 1998: 222; Li & Whaley 2000: 356), which also forms intensive
adjectives with the default /b/ except when the base begins with a labial consonant,
6) One reviewer expressed concern with glossing of the intensive adjectives in (4). As noted
in the book, the dialect Martin (1961) describes is based on the speech of one native
speaker, Peter Onon, who was born in 1919 in Bokore-cien, a place on the Noni River.
They communicated in Japanese. As Street (1963) points out in his review of the book,
there are some inconsistencies in the book due to this fact, one of which is the glosses
on the text. Martin (1961: 1) also attributes some of the facts incompatible with previous
works to dialect mixture and 'standard' influence in Onon's speech at the time.
7) Exceptions are Khlk. bas bat and Klm. bim bitu, mentioned above and to be explained
alternatively below.
314 Hyung-Soo Kim
e.g. Orq. kab kara ‘very black’, Orq. ʃıb ʃıŋarın ‘very yellow’, Orq. kɔb kɔŋnɔrın ~
kɔb kɔŋɔrın ‘very dark’ but Orq. bag bagdarın ‘snow-white’. These examples thus
suggest a constraint at work: the default /b/ is avoided when the base adjective
begins with a labial consonant. Although the number of examples so far gathered
are too small for this generalization to be worked into a full-scale dissimilation rule
for Tungusic, it is due to this constraint a velar consonant /k/ or /g/ appears in
place of the default linking /b/, with the voiced/voiceless variation determined by
voicing agreement with the initial consonant of the following base: /k/ in Xib. faq
farxun but /g/ in Orq. bag bagdarın. Significantly, there is no such compelling
motivation, neither phonological nor morphological, to change the default linking
consonant /b/ to /m/ in Dag. čim čigaan. The claim that /m/ can be a linking
consonant in Dagur is therefore not well founded.8)
It is true that /m/ is used as a linking consonant in modern Turkic languages, e.g.
Tksh. göm gök ‘deep blue’ and Tksh. yem yeşil ‘very green’. In Old Turkic,
however, /p/ was the default linking consonant and use of /m/ as a linker was
restricted to the Oghuz group, the modern descendants of which are Turkish, Azeri,
Turkmen, and Gagauz.9) Note that /b/, the Mongolic congener of Turkic /p/, is also
8) Li & Whaley (2000: 356) posit a rule that inserts /b/ if the first syllable of the adjective
is open, while explaining the /g/ inserted in bag bagdarın ‘very white’ as a full copy of
the first syllable when it is closed. Their rule, however, is problematic in two respects:
1) it fails to note the generalization that the same dissimilation rule occurs in Xib. faq
farxun ‘very dark’ in which the same velar /g/ (with subsequent devoicing to [k] by
voicing assimilation) is inserted because the base also begins with a labial consonant; 2)
/b/ is inserted in kɔb kɔŋnɔrın ‘very black’ even though the first syllable of the base is
closed. The second point needs some explanation as some speakers apparently use
kɔŋɔrın for ‘black’ instead of kɔŋnɔrın but still have the same emphatic prefix. The
difference between these two forms seems to be the suffix -nɔ which appears in most
Tungusic languages, e.g. Evk. kɔŋno-mo, Sol. xoŋnorĩ, and Orq. kɔŋnɔrın but is lacking
in Orq. kɔŋɔrın (cf. Starostin et al. 2003: 720).
9) Note al-Kashgari’s remark in his Compendium: “The rule about colors and exaggerating
the description of things is to take the first letter of the word and join it to bā’ in most
of the Turkic dialects, but to mīm in Oɣuz” (Dankoff & Kelly 1982: 261). This
isoglossic feature of linking consonant more or less continues in modern Oghuz Turkic
languages mentioned here. The few exceptions are in North and North-western Kipchak
languages such as Bashkir, Tatar and Karaim, and Uzbek. These are due either to
influence from the Oghuz Turkic (as in the former languages) or to composite genealogy
(as in Uzbek; cf. Stachowski 2014: 168 and passim, especially chapter 3).